Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
TIMETREE_EMAILNoTimeTree account email. Keep out of source control.
WORKLOG_DB_PATHNoSQLite ledger path. Use :memory: only for tests.<WORKLOG_DATA_DIR>/worklog.sqlite
TIMETREE_MAX_RPSNoMaximum TimeTree requests per second.1
WORKLOG_DATA_DIRNoBase directory for the database, archive, and reports.Platform application-data directory
WORKLOG_TIMEZONENoDefault IANA time zone for dates without an explicit timezone.Africa/Accra
TIMETREE_BASE_URLNoUnofficial TimeTree web base URL.https://timetreeapp.com
TIMETREE_PASSWORDNoTimeTree account password. Keep out of source control.
TIMETREE_PROVIDERNoProvider to use: community or mock.community
WORKLOG_LOG_LEVELNoLog level: debug, info, warn, or error. Logs go to stderr.info
TIMETREE_TIMEOUT_MSNoPer-request timeout in milliseconds.20000
WORKLOG_ARCHIVE_DIRNoCanonical Markdown archive directory.<WORKLOG_DATA_DIR>/worklogs
WORKLOG_REPORTS_DIRNoGenerated progress reports directory.<WORKLOG_DATA_DIR>/reports
TIMETREE_CALENDAR_IDNoThe single private calendar permitted for memos.
TIMETREE_MAX_RETRIESNoRetries for read-only requests only; writes are never retried blindly.3
TIMETREE_REMOTE_MAX_CHARSNoMaximum remote note length. The local archive is never truncated.20000
TIMETREE_MAX_RESPONSE_BYTESNoMaximum accepted upstream response body size in bytes.2000000
TIMETREE_ENABLE_REMOTE_WRITESNoAdditional explicit gate for any remote write.false

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": true
}

Tools

Functions exposed to the LLM to take actions

NameDescription
list_calendarsA

Read-only. Lists calendars available to the configured TimeTree account. Requires local credentials but never writes. Use this to find the one private calendar ID to configure; no destructive calendar operation is exposed.

preview_daily_worklogA

Validate and render a daily work log without saving or contacting TimeTree. Returns the full local Markdown, the exact remote memo title/note that a later publish would send, warnings for suspected secrets, and the previewHash required by publish_worklog. Never invents completion, dates, deadlines, decisions, or evidence.

save_worklogA

Save the work log to the local Markdown archive and SQLite ledger before any remote operation. This tool never contacts TimeTree. Saves are revisioned by date and timezone; identical content is idempotent. Use the returned previewHash with publish_worklog only after reviewing the preview.

get_daily_worklogA

Read a saved work log from the local Markdown archive and SQLite ledger by worklogId or by date plus timezone. This is a local read and does not contact TimeTree.

search_worklogsA

Search the local work-log archive by text, date range, timezone and sync state. Results include matching worklog IDs, dates, sync states and short snippets. This is a local read.

generate_progress_reportA

Build a Markdown progress report from saved local work logs for an explicit date range and timezone. The report attributes every item to its worklog ID and date, distinguishes verified evidence from user-reported statements, and writes a local report file. It is limited to 366 calendar days per request and does not invent missing work, decisions, blockers, deadlines or completion.

publish_worklogA

Explicit remote write. Publishes a saved work log as one all-day TimeTree memo in the single configured private calendar. Requires remote writes enabled, credentials, a previewHash that matches the current saved revision, no suspected secrets, and confirmation PUBLISH. It reconciles first to avoid duplicates, verifies the created memo by reading it back, and returns UNCERTAIN or CONFLICT instead of claiming success. No destructive or update operation is provided.

get_sync_statusA

Read local synchronization state, counts, attempt history and provider health. Without a worklogId or date it returns an aggregate summary. It never writes and does not call TimeTree.

retry_pending_syncA

Reconcile pending, failed or uncertain work logs before any retry. Defaults to dryRun: true; a dry run records a verified memo or conflict but does not persist an empty listing as absence. A live retry requires remote writes enabled and confirmation RETRY, is limited to 10 work logs per call, never retries CONFLICT, FAILED_PERMANENT, BLOCKED or NOT_PUBLISHED records, and never performs a first publication. Re-verifies every created memo. Never assume a timeout means no memo.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A4/5.0

Scored across 9 tools

Disambiguation4/5

Each tool maps to a distinct stage of the worklog lifecycle (preview, save, get, search, publish, sync status, retry), and descriptions clarify the read-only vs remote-write boundaries. Minor overlap between preview_daily_worklog and save_worklog, and between get_sync_status and search_worklogs' sync-state filtering, but the descriptions disambiguate well.

Naming Consistency4/5

All tools use snake_case verb_noun patterns consistently (list_calendars, save_worklog, publish_worklog, retry_pending_sync). Minor deviation: 'daily' appears in preview_daily_worklog and get_daily_worklog but not in save_worklog, slightly breaking the resource-naming symmetry.

Tool Count5/5

Nine tools form a tight, well-scoped set covering the full local-to-remote worklog flow. Every tool has a clear, non-redundant role with no filler.

Completeness4/5

Covers calendar discovery, local preview/save/read/search, reporting, publishing, sync status and retry — strong lifecycle coverage. Gaps exist (no delete/archive removal for local worklogs, no unpublish or update of published memos), but these appear to be intentional safety omissions rather than dead ends.